JOURNAL OF APPLIED SCIENCES RESEARCH

Copyright © 2014, American-Eurasian Network for Scientific Information publisher
JOURNAL OF APPLIED SCIENCES RESEARCH
JOURNAL home page: http://www.aensiweb.com/JASR
2014 Special; 10(15): pages 8-18.
Published Online 6 December 2014.
Research Article
What clues can we get from a student recruitment website? An application of
web analytics
1Yung-Sheng
Yang, 2Chiang-Yu Cheng, 3Jui-Hsien Shih, 4Su-Shiang Lee
1
Ph.D. Student, College of Management, Chaoyang University of Technology, Taiwan.
Assistance Professor, Department of Marketing and Logistics Management, Chaoyang University of Technology, Taiwan.
Ph.D. Student, College of Management, Chaoyang University of Technology, Taiwan.
4
Professor, Dean of Department of Management, Chaoyang University of Technology, Taiwan.
2
3
Received: 30 September 2014; Revised: 17 November, 2014; Accepted: 25 November, 2014; Available online: 6 December 2014
© 2014 AENSI PUBLISHER All rights reserved
ABSTRACT
Nearly all universities worldwide operate a student recruitment website because potential students access registration announcements
from different regions and countries. However, tracking visitor footprint on such an information-oriented website should be more critical
than maintaining a website merely for information diffusion. The more knowledge website owners possess regarding visitors’ online
activities, the more able they are to improve website performance. We applied web analytics to uncover the flow secret hidden in student
recruitment websites and demonstrate the performance metrics that contribute to website success.
Key words: Authoring Tools and Methods; Country-specific Developments; Navigation;
INTRODUCTION
Online visitors leave “footprints” each time they
visit a website. The number of footprints on
websites, multiplied by the number of websites
worldwide, represents an abundance of online
behavior that researchers call “big data” [7], which
serves as an important source of business
intelligence. For example, a retailer with a big data
adoption is expected to increase its operating margin
by more than 60%, whereas services enabled by
broad location-based data allow companies to earn
over $600 billion US dollars in revenue [18].
However, White (2012) argued that big data is
difficult to quantify, using traditional data processing
applications (e.g., on-hand database management
tools) to manage such large and complex data. The
potential of big data can only be realized using tools
with the functions of capture, storage, search,
sharing, transfer, analysis, and visualization [26] to
process it. Web analytics is an outstanding tool that
can be used to analyze big data and has received
wide attention from current practices. Bloomberg
Business Week (2011) reported that 97% of highgrowth companies (with revenues over $100 million
US dollars in the survey year) were observed to
apply web analytics to their business operations.
Gartner [11] highlighted web analytics as one of the
10 strategic technology trends for 2013.
Researchers have applied web analytics to
analyze several aspects of online behavior, including
[21], identifying consumer goals [19], and online
decision-making processes [5]. However, the
applicability of web analytics to a student
recruitment website has received scant attention
from scholars. Hossler [13] suggested that
recruitment agencies at universities must actively
evaluate whether their websites convey the type of
information potential students require, rather than
merely decorating their websites with unnecessary
content that students may not require. Although
universities often take advantage of recruitment
websites to communicate large amounts of
information widely and rapidly to potential students
[1], the effectiveness of using a recruitment website
remains unclear. Because universities currently face
increased competition with numerous other
competitors, recruitment agencies must focus their
attention on web analytics to facilitate recruitment
practices. We introduce the applicability of web
analytics to student recruitment websites.
Anomalous states of knowledge [3] pertains to why
potential students are likely to visit a recruitment
website during their school selection processes,
whereas web analytics detects clues regarding
students’ website-visiting behavior. We answer the
following research questions:
RQ1. How popular is the student recruitment
website?
RQ2. How do visitors behave on the student
recruitment website?
RQ3. What traffic-sources exist in the student
Corresponding Author: Chiang-Yu Cheng, Department of Marketing and Logistics Management, Chaoyang University of
Technology, Taiwan.
Tel: +886-423323000, Fax: +886-423742369, E-mail: [email protected]
9
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
recruitment website?
RQ4. What devices do visitors use to visit the
recruitment website?
RQ5. Which location do the visits originate
from?
RQ6. Does web analytics sufficiently overcome
recruiting challenges?
Our outcomes contribute to the relevant
literature by discovering theoretical phenomena and
offering managerial implications to help both
researchers and recruitment agencies to monitor
students’ information requirements during their
visits to recruitment websites. Our study is one of the
first to show the applicability of web analytics in the
technology and education context. Thus, the findings
are beneficial to universities worldwide. This paper
is organized as follows. Section 2 presents a
description of the theoretical foundation of this
study, followed by a literature review in Section 3.
Section 4 introduces the research method and
Section 5 provides the data analysis and results.
Section 6 presents a discussion on the research
findings and implications. Finally, Sections 7 and 8
offer a discussion of this study and present
limitations.
2. Theoretical Background:
The primary purpose of building a student
recruitment website is to promote the university and
recruit new students. However, the success of such
an information-oriented website is dependent on why
potential students visit that website. Belkin [3]
proposed ASK to help answer this question.
According to ASK, a person whose knowledge is
insufficient for solving a confronted problem
(anomalous state) conducts information searching
for the required information, which can be
summarized in three steps, as follows:
(a) A person perceives that his or her knowledge is
insufficient for solving a specific problem, thus
requiring additional information searching.
(b) The person then requests information from a
certain source (e.g., magazine, friend, Internet),
which returns the requested information.
(c) The person then makes a judgment regarding the
received information and determines the degree
of fulfillment in accommodating her or his
knowledge gap. The greater the fulfillment is, the
more likely that this person will cease
information searching. Conversely, the person
who perceives that the received information is
insufficient for solving the problem might
iteratively conduct information searching and
request the required information from the same
or different sources.
Iterative information searching in the final step
implies that a person’s knowledge state relies on
fulfillment and is therefore dynamic, rather than
fixed [4]. A potential student who seeks admission to
a university might initially screen recruitment
websites of various universities and form a shortlist
that represents his or her consideration set. Because
the set consists of self-selected options, information
presented on these alternatives may be more
attractive than those that are excluded from the set.
Consequently, students frequently take advantage of
received information to narrow their consideration
sets, particularly when they discover numerous
options, similar to that of product selection in
consumer shopping [23]. However, additional
information searching is necessary if students have
little knowledge regarding the universities in their
consideration sets (i.e., knowledge regarding school
selection remains unclear or anomalous) and such
information searching will not end until they are
satisfied that they possess sufficient school-selection
knowledge for making a final choice (i.e., needs
fulfillment). Therefore, students rely on information
that recruitment websites provide to alleviate their
ASK (or uncertainty) regarding school selection
[10].
3. Literature Review:
3.1. Student Recruitment Challenges:
Student
recruitment
agencies
frequently
experience challenges when endeavoring to persuade
potential students. Ross, Heaney, and Cooper [22]
emphasized the challenge of limited marketing
budgets, in which school administrators must seek
institutional readiness (e.g., marketing department
size, employee qualifications, institutional recruiting
experience, and institutional focus) to manage
limited educational resources and to reduce
marketing costs. Another challenge is that different
objectives may require various recruitment methods
and strategies. However, methods and strategies that
are valid in an education sector are not necessarily
feasible in other sectors [22]. The same strategy may
not be viable for different or even similar education
sectors. Anctil [1] identified one of the challenges
for recruitment agencies as differentiating
themselves from competitors because of the
necessity of abundant evidence (e.g., certification,
faculty performance, and learning environment) of
the school’s reputation. Although this physical
evidence plays an indispensable role in student
persuasion, its effectiveness in student recruitment
remains unclear. Lindbeck and Fodrey [16] indicated
another challenge from a different perspective. They
argued that admission departments throughout the
United States are enthusiastically engaged in using
technology in student recruitment activities;
however, few staffs have confidence they are fully
benefiting from its adoption. Because scholars and
practitioners regard marketing cost, differentiation,
and confidence of technology assistance as crucial to
student recruitment, we apply web analytics to
10
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
address these critical issues.
3.2. Web Analytics:
Web analytics is a monitoring technique that
collects, measures, analyzes, and reports on Internet
data to elucidate visitors’ online behavior [8]. Web
analytics can be categorized into two types: off-site
and on-site. Off-site analytics can be used regardless
of whether the analyst owns a website, whereas onsite analytics can only be used if the analyst owns a
website or has the permission to access a website.
Off-site web analytics primarily focuses on website
opportunities (potential visitor), visibility (the
number of registered members), and visitor
comments (word-of-mouth), whereas on-site web
analytics measures visitor behavior on a specific
website for activities that off-site web analytics do
not address (the number of website visitors who
conduct a specific action beyond a casual content
view). We introduce web analytics applicability to a
student recruitment website. On-site web analytics is
more applicable than off-site web analytics for three
reasons: (a) nearly all universities operate a selfmaintained student recruitment website, which
means that universities embed on-site web analytics
codes into their websites without requesting access
permission from others; (b) recruitment agencies use
the student recruitment website to persuade potential
students, and on-site web analytics helps them to
estimate how those students interact with the website
by monitoring what pages potential students have
visited, where they were referred from, how much
time they spent on a website, and their clicking
behaviors on the website; and (c) because students
may exclude a recruitment website from their
consideration sets if that website provides no clear
and required recruitment information, it is necessary
to narrow the gap between the amount of
information required to perform school selection and
the amount of information potential students already
possess. On-site web analytics assists recruitment
agencies in understanding the actual online behavior
of potential students to improve student needs for
recruitment information. For instance, on-site web
analytics can calculate the bounce rate for a
recruitment website (the percentage of visitors who
immediately enter and leave that website), indicating
whether the website entry page causes potential
students to discontinue viewing the website without
viewing other pages. Because cost-per-student
recruitment is critical for any recruitment agency
[27], on-site web analytics not only addresses the
ASK issue, but also contributes to the effectiveness
of a student recruitment website.
4. Research Method:
4.1. Website Profile:
Chaoyang University of Technology (CYUT) is
the first private technology university in Taiwan. It
announced a new cooperative education program
(also called the dual-track program) in the spring of
2013 to enroll new students. The recruitment team at
CYUT was assigned to create a recruitment website
(http://www.cyut.edu.tw/-ccy)
for
presenting
enrollment information to potential students,
including
program
introduction,
application
materials, admission quota, admission timeline,
department introduction, tuition fee, and traffic
information (Fig. 1). The promotion for this dualtrack program was conducted from May 15 to June 5
in 2013 and the web analytic codes were removed
afterward. The success of this recruitment website
relied on the number of visitors who downloaded
application materials during their website visit.
Fig. 1: The Recruitment Website of Cooperative Education Program at CYUT
11
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
4.2. Google Analytics:
Web analytics tools are prolific, ranging from
clickstream to competitive intelligence; certain
among them are free (e.g., WebTrends, Yahoo! Web
Analytics, Google Analytics), and others are highly
expensive
(e.g.,
Adobe
Analytics,
IBM
Coremetrics). Because CYUT wished to monitor the
visiting behavior of potential students on the
recruitment website with minimum cost, a free web
analytics tool was a necessary consideration. The
web analytics tool with a user-friendly interface
(e.g., Graphical User Interface, GUI) and multilanguage support is more appropriate than one that
provides only a command line interface and English
language support. Finally, the ease of implementing
a tracking code, which is used to connect the thirdparty analytic server, should also be considered.
Although nearly all web analytics tools include a set
Table 1: The Key Metrics Used in the Study
Research Question
Key Metric
Visits
How popular is the student recruitment
website?
Visitors
% New visit
Unique visitors
Bounce rate
Page views
Page/visit
How do visitors behave on the student
recruitment website?
Avg. visit duration
Total events
Unique events
What are the sources of traffic found in the
student recruitment website?
What devices visitors use to visit the
recruitment website? Where do they come
from?
Search traffic
(keyword)
Referral traffic
Direct traffic
(landing page)
Technology
(Browser, OS, devices)
Demographics
(location)
5. Data Analysis:
5.1 How Popular Is the Student Recruitment
Website?:
Figure 2 shows the popularity of the recruitment
website. There were 2,015 new visitors (59.5%) and
1,371 (40.5%) returning visitors, constituting a total
number of 3,386 visits, whereas 2,024 of them were
unique visitors. Visitors who had never visited the
recruitment website before initiated 59.33% of new
of basic metrics, the CYUT recruitment team
adopted Google Analytics (GA) because of its free
service, detailed statistics about visitors to the site,
user-friendly GUI, and multi-language support. The
ease of use, in particular, makes GA one of the most
popular web analytics tools worldwide; according to
a survey conducted by TechCrunch [24], GA is the
most widely used web analytics tool and is currently
used by approximately 55% of the 10,000 most
popular websites.
4.3. Key Metrics:
Table 1 presents a summary of the key metrics
used in this study. All the metrics were adopted from
Google
Conversion
University
(http://www.google.com/analytics/iq.html) and were
used to answer the proposed research questions
listed in Section 1.
Description
The number of single visits initiated by all the visitors to
the website. If a user is inactive on the website for over
30 minutes, any future activity will be counted to a new
visit.
The number of visitors who have “ever” visited the
website for a date range.
An estimate of the percentage of first time visits.
The number of unduplicated visitors to the website over
the course of a specified period time
The percentage of single page visits in which the person
left the website from the entrance page without
interacting with the page.
The total number of pages viewed.
The average number of pages viewed during a visit to
the website
The average time duration of a visit.
The number of times events occurred.
The number of visits during which one or more events
occurred.
What is the most popular search engine that the visitors
used? (What keywords are they used?)
Where does the referral traffic come from?
How many visitors directly type the URL to get into the
website? (The web page that visitors arrive at after they
type the URL)
Technology is one of dimension that monitors what
browsers, operating systems, and devices that visitors
use to visit the website.
Location is one of dimensions that records where do the
visitors come from.
visits. Although this recruitment website seems to
receive many visitors, the metrics of visits, visitors,
and the percentage of new visits may be
overestimated without accounting for the bounce
rate. As shown in the same figure, 43.18% of visitors
entered the website but exited immediately without
visiting any other pages. The visit trend curve shows
that nearly all rush flows on the peaks appeared at
night, with the exception of June 5.
12
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
Fig. 2: A Snapshot of Google Analytics from the Recruitment Website
5.2 How Do Visitors Behave on the Student
Recruitment Website?
Visitors viewed 8,812 pages during the
promotion period, and viewed 2.6 pages were per
visit (Fig. 2). The low pages-per-visit count (fewer
than four pages in average) indicates that visitors
come to the recruitment website but do not want to
stay. This phenomenon is consistent with the finding
in Fig. 2, that visitors spent only 4 min and 34 s on
the recruitment website on average. Because the
recruitment website is a functional website that
conveys necessary information for potential students,
an average student may view only two or three pages
and stay for slightly over a minute if the website
information fulfills their visiting purpose (e.g.,
application materials download, dual-track program
introduction). This assertion is demonstrated by the
data shown in Table 2, that the number of
application material downloads was 991 (total
events) and the number of visits during which the
application materials were downloaded was 768
(unique events). Therefore, the visitors on this
recruitment website were considerably purposeful.
Table 2: The Number of Total Events and the Unique Events of the Recruitment Website
5.3. What Traffic-sources Exist for the Student
Recruitment Website?:
Figure 3 illustrates the traffic sources of the
student recruitment website. Direct traffic was the
most common method of entering the website (2,762
visits, 81.6%) followed by referral traffic (573 visits,
16.9%), and search traffic (51 visits, 1.5%). This
implies a two-fold implication. The in-class
recruiting activities and recruiting publications are
relatively appealing because most visitors directly
type the URL they received from those information
sources to access the recruitment website. However,
the affiliation of the recruitment website with other
websites must be further strengthened because both
referral traffic and search traffic was lower than was
direct traffic.
13
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
Fig. 3: Traffic Sources of the Student Recruitment Website
5.4. What Browsers and Devices Do Visitors Use to
Visit the Recruitment Website?:
Table 3 shows that both Chrome (1,467 visits,
43.33%) and Internet Explorer (1,149 visits, 33.93%)
are the most popular browsers among visitors.
Android Browser (438 visits, 12.94%) and Safari
(182 visits, 5.38%) ranked as the third- and fourthtier browsers that visitors used.
Table 3: Browser Usage for the Visits of the Student Recruitment Website
Regarding operating system usage (see Table 4),
Windows (2,664 visits, 78.68%) remains the
prevailing operating system that visitors adopt.
However, mobile data access is a new phenomenon
in web analytics; Android (547 visits, 16.15%) and
iOS (147 visits, 4.34%) combined contribute 20.49%
of distribution among all types of operating systems.
Table 4: The Distribution of Operating System Usage
Although the student recruitment website
showed that mobile data access is becoming
increasingly popular, a mobile device can be either a
tablet or a smartphone, each of which has different
screen sizes and resolutions that must be further
identified. Table 5 presents a summary of the screen
resolution of mobile devices that visitors used. Of
the 694 mobile visitors, 73 used a larger screen
resolution (over 7 in － 1280  800) to visit the
recruitment website, whereas 621 used a smaller
screen resolution (under 5 in－1280  720) as their
visiting device. This finding reveals a need for the
student recruitment website to provide visitors with
multi-resolution support so that they are able to visit
14
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
the website without constantly changing (magnifying
or minifying) their screen size.
Table 5: The Distribution of Mobile Device Usage
5.5 Which Location Do Visitors Come from?:
Knowing where the visitors come from
provides valuable insight for recruiting activities
because recruitment agencies can rely on such clues
to evaluate the success of their recruiting activities.
For example, in-class recruiting activities are
considered more successful in one location than in
other locations if most visitors come from that
location, but not from others. Table 6 outlines where
the recruitment website visitors originated from. Of
the 3,374 visitors (12 visitors were excluded from
the analysis because they were located outside of
Taiwan), most visitors came from Central Taiwan
(i.e., locations 1 － 10), whereas relatively few
visitors came from other locations. Thus, the
recruiting activities in these low-visitor locations
should be enhanced.
Table 6: The Distribution of Visitor Locations
5.6. Psychological Responses after Using Web
Analytics:
We applied the constructs of satisfaction and
continuance intention to measure agencies’
psychological responses after using web analytics for
the student-recruitment website. Satisfaction is the
extent to which recruitment agencies believe the web
analytics available to them meet their information
requirements, such as web traffic reports [14],
whereas continuance intention is the intention of
recruitment agencies to continue using web analytics
[15]. We measured both constructs using a 5-point
Likert scale adapted from Park, Kim, and Koh [20],
where 1 = strongly disagree and 5 = strongly agree.
We created a web-based questionnaire for the data
collection. A total of 47 recruitment agencies were
requested to participate in this survey. Prior to the
formal analysis, we conducted a preliminary data
examination, screening for missing data, outliers,
construct reliability, and construct validity. No
missing data or outliers were observed. Reliability
was evaluated using composite values. Hair, Black,
Babin, Anderson, and Tatham [12] recommended an
acceptance level of 0.7 for composite reliability. As
shown in Table 7, the composite reliabilities of
satisfaction and continuance intention constructs
exceeded 0.88, meeting this criterion. Fornell and
Larcker [9] suggested two criteria to establish
15
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
convergent validity. First, all factor loadings should
be significant and exceed 0.5. Second, the average
variance extracted (AVE) for each construct should
exceed the measurement error variance for that
construct (AVE should be greater than 0.5). All the
items listed in Table 7 exhibit loadings greater than
0.80 within their respective constructs and all AVEs
are larger than the error variance. Thus, both criteria
for convergent validity were met. Discriminant
validity is the extent to which a construct and its
indicator variables differ from another construct and
its indicator variables [2]. We examined it using a
criterion suggested by Fornell and Larcker [9]: the
square root of AVEs should be greater than the
correlation between the two constructs. Table 8
shows that the correlation between the pair of
constructs was less than the corresponding AVEs
(diagonal values). All the constructs met the
requirement, providing evidence of discriminant
validity.
Table 7: Summary of Measurement Scales
Construct
Measure
Satisfaction (SAT)
composite reliability = 0.92
SAT1
After using web analytics, I am satisfied with the reports it gave to me
SAT2
After using web analytics, I am satisfied with the implications derived
from the reports
SAT3
Overall, I am satisfied with the assistance given by web analytics
Continuance intention (CINT)
composite reliability = 0.88
CINT1
CINT2
CINT3
Factor Loading
0.86
0.93
0.90
I will use this web analytics next time when I am assigned to participate
0.82
in student recruitment activities
I do not consider any alternative web analytics next time when I am
0.80
assigned to participate in student recruitment activities
I intend to recommend this web analytics to other colleagues every time
0.90
when they use a website to be an information disseminator
Table 8: Correlation and AVE
Construct
AVE
SAT
CINT
Satisfaction (SAT)
0.80
0.89
Continuance intention (CINT)
0.71
0.49
0.84
*Diagonal elements in bold are the square root values of the average variance extracted (AVE). Off-diagonal element is the correlation
between the two constructs.
Figure 4 shows the ratings and distributions of
the research constructs. The results revealed that
nearly all recruitment agencies were satisfied or
strongly satisfied with the assistance given by web
analytics. They also expressed a strong willingness
to continue using web analytics for future
assignments to participate in student recruitment
activities.
Fig. 4: Ratings and Distributions of the Research Constructs
To confirm the causality between satisfaction
and continuance intention, we conducted partial least
squares (PLS) analysis (Fig. 5). The result indicated
that continuance intention is a function of
satisfaction; satisfaction is a strong predictor of
continuance intention because it explained 71%
variance of continuance intention (R2 = 0.71).
16
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
β=0.84, t=20.08
p<0.001
Satisfaction
Discussion:
Our findings provide insights into the
challenges of student recruitment activities,
particularly in the aspects of marketing cost,
differentiation, and confidence of technology
assistance. The following outlines how web analytics
can help (a) reduce marketing cost, (b) create
different recruiting activities, and (c) increase
agencies’ confidence in technology assistance.
6.1. Marketing Cost:
The analysis identified that nearly all rush
flows on the visit trend curve occurred at night (Fig.
2). This phenomenon does not indicate that CYUT
must ask their employees to work overtime; instead,
it produces a technology-enabled opportunity for
CYUT to interact with their nighttime visitors
without time limitations. This can be archived by
embedding an instant messaging service (IMS) into
the student recruitment website. With the aid of IMS,
nighttime visitors can ask questions and contact staff.
They can also leave messages and obtain responses
from the IMS robot late at night when staff members
are not online. In addition, CYUT frequently
dispatches recruiting teams (e.g., professors) to
senior high schools for program promotion. CYUT
believes that personal in-class interaction is the most
effective method of recruiting students successfully.
However, this approach is costly (e.g., travel
allowances, accommodation) and makes random
attempts at recruiting activities (i.e., it is unknown
whether potential students will actually visit the
recruitment website after joining the recruitment
meeting). The traffic source report provided by web
analytics clearly indicates where website visits
originate from and therefore recruitment agencies
can specifically understand which locations must be
visited. For example, Section 5.5 shows that most
website visits originate from cities located in Central
Taiwan, implying the necessity of promotional
efforts in other locations and thus brings marketing
budgets to bear on the right location. However, from
the analysis shown in Fig. 3, recruitment agencies
can learn that search engines do not contribute
sufficient traffic to the recruitment website. Thus, we
suggest adopting search engine optimization (SEO)
to attract more visitors, such as keyword advertising,
to help CYUT protect its marketing budget.
6.2. Differentiation:
In contrast to traditional marketing campaigns,
R2=0.71
Continuance
intention
mobile marketing focuses primarily on consumers
who tend to be more dynamic. Similarly, potential
students use their mobile devices to obtain necessary
information from a school website. According to the
results in Table 4, approximately 20% of visitors
used mobile devices to visit the recruitment website.
This number is expected to increase in the near
future because of the pervasiveness of mobile
devices. We therefore suggest that CYUT consider
expanding its mobile advertising efforts when web
analytics reports show consistent increases in mobile
traffic to the recruitment website. CYUT should also
create a mobile-friendly recruitment website; if
visitors perceive that the recruitment website is not
mobile-friendly, they are highly likely to leave the
website without an in-depth visit, which results in a
high bounce rate. Thus, capitalizing on traffic from
mobile visitors is imperative. Knowledge of mobile
access to the website and providing a mobile version
of the website are viable methods for CYUT to
differentiate itself from competitors.
6.3. Confidence of Technology Assistance:
Because the dual track program is an annually
announced program and user dissatisfaction with a
system leads to discontinued use [17], maintaining
user satisfaction toward a system is pivotal. Figures
4 and 5 confirm this assertion; only one agency was
dissatisfied with the assistance provided by web
analytics, whereas the others were satisfied or highly
satisfied with the benefits of using web analytics.
Most agencies indicated their desire to use web
analytics for future assignments in a recruitment
team. The greater the satisfaction recruitment
agencies perceive, the more likely they are to
recommend web analytics to others or use it
continually. In other words, agencies’ satisfaction
after using web analytics for a recruitment website
depends on report and implication satisfaction,
meaning that the clues and insights recruitment
agencies obtain from the reports are crucial for
satisfaction. In our study, GA provided reports,
including visitor segmentation based on visiting
behavior on the website, measurement and
monitoring of website traffic, monitoring external
referrers, and monitoring clickstreams. These
insightful reports in turn elicit clues and implications
of visitor website behavior to enable the CYUT
recruitment website to reach the desired visitors with
minimal marketing expenditure. Measuring and
monitoring website traffic also helps to improve the
17
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
recruitment website for future applications.
Monitoring external referrers helps to identify which
search engines or affiliated websites contribute most
to website traffic, and monitoring clickstreams
identifies unvisited and poorly performing web
pages to help maintain a successful recruitment
website. These clues and implications confirm
agency satisfaction with web analytics; they can be
used to predict continued use of web analytics.
Conclusion:
The student recruitment website is an
information disseminator that conveys necessary
information to potential students, who can
immediately enter and leave the website without
viewing additional pages. They can visit the website
and browse it until they believe they have obtained
sufficient information. They can also visit the
website each time ASK occurs. Regardless of the
type of visiting behavior, web analytics elucidates
website traffic. Private school enrollment has
dropped considerably in the past decade [25];
therefore, it is vital to apply certain approaches to
prevent universities from becoming a victim of the
recession. Web analytics in this aspect demonstrates
outstanding performance. After implementing the
tracking code on the recruitment website,
universities can effortlessly analyze the visiting
behavior of potential students (e.g., average page
visit duration, most commonly downloaded files,
navigation paths). They can also monitor mobile
traffic to revamp the current version of their
recruitment websites because numerous potential
students currently visit the website on a mobile
device. In sum, many universities suffer as result of
an aging population and low birthrate; it is inevitable
that they must compete for a decreasing number of
applicants. CYUT is one of a small number of
private universities in Taiwan that have successfully
applied web analytics to student recruitment
websites, and is experiencing increasing enrollment
despite the low birthrate. Speculation on how CYUT
has established itself in an unassailable position is
that it applies technology to pedagogy (e.g., digital
black board system, distance teaching system) and to
student admission, student enrollment, and most
crucially, to student recruitment.
8. Limitations:
This study has certain limitations. First, we
demonstrated the applicability of web analytics to
the student recruitment website, however, other tools
exist (e.g., Adobe Site Catalyst, IBM Coremetrics),
many of which possess characteristics and
capabilities beyond GA. Future studies could
investigate the advantages and disadvantages of
these analytic tools and compare them with the one
used in this study. Second, GA uses cookies stored
in the user’s computer to track the number of
website visits. However, when the cookie expires,
GA does not count the cookie owner as a returning
visitor, but as a new visitor; GA may overestimate or
underestimate in either case. Those using GA to
monitor website traffic should carefully evaluate the
number of visits in this regard. Third, GA is unable
to detect net spiders (automatic visiting robots). The
number of website visits may be inflated when a
robot visits the tracking website. One way to solve
this problem is to add “allow: robot name” or
“disallow: robot name” syntax in the robot.txt, which
is embedded in the tracking website. Finally, the
number of visits is inaccurate if the tracking website
receives extremely low traffic. GA must evaluate
numerous visits to enable the analysis results to
reflect accurate visitor behavior; otherwise, it will
result in an inaccurate view of user interaction with
the website.
References
1.
Anctil, E.J., 2008. Selling Higher Education:
Marketing and Advertising. ASHE Higher
Education Report., 34: 1-121.
2. Bagozzi, R., L. Phillips, 1991. Assessing
Construct Validity in Organizational Research.
Administrative Science Quarterly, 36(3): 421458.
3. Belkin, N.J., 1980. Anomalous States of
Knowledge as a Basis for Information Retrieval.
The Canadian Journal of Information Science, 5:
133-143.
4. Belkin, N.J., H.M. Brooks, R.N. Oddy, 1982.
Ask for Information Retrieval. Journal of
Documentation, 38: 61-71.
5. Bhatnagar, A., 2009. Web Analytics for
Business Intelligence beyond Hits and Sessions.
Online., 333(6): 32-35.
6. Bloomberg, 2011. The Current State of Business
Analytics: Where Do We Go from Here? 2011.
http://www.sas.com/resources/asset/busanalytics
study_wp_08232011.pdf (accessed September
17, 2013).
7. Chen, H.C., R.H.L. Chiang, V.C. Storey, 2012.
Business Intelligence and Analytics: From Big
Data to Big Impact. MIS Quarterly, 36(4): 11651188.
8. Clifton Brian, 2012. Advanced Web Metrics
with Google Analytics. Wiley.com.
9. Fornell, C., D.F. Larcker, 1981. Evaluating
Structural Equation Models with Unobservable
and Measurement Error. Journal of Marketing,
18(1): 39-50.
10. Galbraith, J.R., 1974. Organization Design: An
Information Processing View. Interfaces. 4(3):
28-36.
11. Gartner. Top 10 Strategic Technology Trends
for 2013. 2012. http://apmdigest.com/gartnertop-10-strategic-technology-trends-for-2013big-data-cloud-analytics-and-mobile (accessed
September 17, 2013).
18
Chiang-Yu Cheng et al, 2014 /Journal Of Applied Sciences Research 10(15), Special, Pages: 8-18
12. Hair, J.F., C.B. William, J.B. Barry, E.A. Rolph,
2009. Multivariate Data Analysis: A Global
Perspective. 7th ed. Upper Saddle River:
Prentice Hall.
13. Hossler, D., 1999. Using the Internet in College
Admission: Strategic Choices. Journal of
College Admission.,162: 12-18.
14. Lee, M.C., 2010. Explaining and Predicting
Users’ Continuance Intention Toward Elearning: An Extension of the Expectationconfirmation Model. Computers & Education.
54(2): 506-516.
15. Lin, W.S., C.H. Wang, 2012. Antecedents to
Continued Intentions of Adopting E-learning
System in Blended Learning Instruction: A
Contingency Framework Based on Models of
Information System Success and Tasktechnology Fit. Computers & Education., 58:
88-89.
16. Lindbeck, R., B. Fodrey, 2010. Using
Technology in Undergraduate Admission: A
Student Perspective. Journal of College
Admission., 208: 10-17.
17. Oliver, R.L., 1980. A Cognitive Model of the
Antecedents and Consequences of Satisfaction
Decisions. Journal of Marketing Research,
17(4): 460-469.
18. McKinsey. Big Data: The Next Frontier for
Innovation, Competition, and Productivity,
2011.
http://www.mckinsey.com/insights/business_tec
hnology/big_data_the_next_frontier_for_innova
tion (accessed September 19, 2013).
19. Pakkala, H., K. Presser, T. Christensen, 2012.
Using Google Analytics to Measure Visitor
Statistics: The Case of Food Composition
Websites. International Journal of Information
and Management., 32(6): 504-512.
20. Park, J.S., J.J. Kim, J. Koh, 2010. Determinants
of Continuous Usage Intention in Web
Analytics Services. Electronic Commerce and
Research Applications., 9(1): 61-72.
21. Plaza, B., 2011. Google Analytics for
Measuring Website Performance. Tourism
Management., 32(3): 477-481.
22. Ross, M., J.G. Heaney, M. Cooper, 2007.
Institutional and Managerial Factors Affecting
International Student Recruitment Management.
International
Journal
of
Educational
Management., 21(7): 593-605.
23. Shocker, A.D., M. Ben-Akiva, B. Boccara,
1991. Consideration Set Influences on
Consumer Decision-making and Choice: Issues,
Models, and Suggestion. Marketing Letters,
2(3): 181-197.
24. TechCrunch. Google Biz Chief: Over 10M
Websites Now Using Google Analytics, 2012.
http://techcrunch.com/2012/04/12/googleanalytics-officially-at-10m (accessed September
22, 2013).
25. Census Bureau, U.S., 2013. The Decline in
Private
School
Enrollment.
2013.
http://www.census.gov/hhes/school/files/ewert_
private_school_enrollment.pdf.
26. Vance A. Start-up Goes after Big Data with
Hadoop
Helper,
2010.
http://bits.blogs.nytimes.com/2010/04/22/startup-goes-after-big-data-with-hadoop-helper.
27. Warschauer, M., 2010. Invited Commentary:
New Tools for Teaching Writing. Language
Learning & Technology, 14(1): 3-8.
28. White, T., 2012. Hadoop: The Definitive Guide.
O’Reilly.